, Volume 23, Issue 3, pp 281–302 | Cite as

A modular integration and multiresolution framework for image interpretation

  • K Sunil Kumar
  • U B Desai


In this paper, we give a generalized formulation for a vision problem in the framework of modular integration and multiresolution. The developed framework is used to solve the high-level vision problem of scene interpretation. The formulation essentially involves the concept of reductionism and multiresolution, where the given vision taskν is broken down into simpler subtasksν 1,ν 2, …,ν m. Moreover, instead of solving the vision taskν Ω =ν at the finest resolution Θ, we solve the synergetically coupled vision subtasks at coarser resolutions (Ω −N) for Θ ≥N>0 and use the results obtained at resolution (Θ −N) to solveν Ω −N+1, the vision task at resolution (Θ −N+1). Image interpretation is a two-phased analysis problem where some salient features or objects in an image are identified by segmenting the image and the objects in the segmented image are interpreted based on their spatial relationships. We present a solution to the joint segmentation and interpretation problem in the proposed generalized framework. For the interpretation part we exploit the Markov Random Field (MRF) based image interpretation scheme developed by Modestino and Zhang. Experimental results on both indoor and outdoor images are presented to validate the proposed framework.


Modular integration multiresolution framework image interpretation Markov random field 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Binford T 1982 Survey of model based image analysis systems.Int. J. Robotics Res. 1: 587–633Google Scholar
  2. Bozma H I, Duncan J S 1994 A game-theoretic approach to integration of modules.IEEE Trans. Pattern Anal. Machine Intell. 16: 1074–1086CrossRefGoogle Scholar
  3. Chu C C, Aggarwal J K 1992 Image interpretation using multiple sensing modalities.IEEE Trans. Pattern Anal. Machine Intell. 14: 840–847CrossRefGoogle Scholar
  4. Clark J J, Yullie A L 1990Data fusion for sensory information processing systems (Norwell, MA: Kluwer)Google Scholar
  5. Clement V, Thonnat M 1993 A knowledge-based approach to integration of image processing procedures.CVGIP: Image Understanding 55: 166–184CrossRefGoogle Scholar
  6. Gamble E B, Geiger D, Poggio T A, Weinshall D 1989 Integration of vision modules and labelling of surface discontinuities.IEEE Trans. Syst., Man Cybern. 19: 1576–1581CrossRefGoogle Scholar
  7. Geman S, Geman D 1984 Stochastic relaxation, Gibbs distribution, and Bayesian restoration of images.IEEE Trans. Pattern Anal. Machine Intell. 6: 721–741MATHCrossRefGoogle Scholar
  8. Jeng 1992 Subsampling of Markov random fields.J. Visual Commun. Image Representation 3: 225–229CrossRefGoogle Scholar
  9. Jolion J M, Rosenfeld A 1994A pyramid framework for early vision (Norwell, MA: Kluwer)Google Scholar
  10. Kim I Y, Yang H S 1993 Efficient image labelling based on Markov random field and error backpropagation network.Pattern Recogn. 26: 1695–1707CrossRefGoogle Scholar
  11. Kim I Y, Yang H S 1995 An integrated approach for scene understanding based on Markov random field.Pattern Recogn. 28: 1887–1897CrossRefGoogle Scholar
  12. Kumar V P, Desai U B 1996 Image interpretation using Bayesian networks.IEEE Trans. Pattern Anal. Machine Intell. 18: 74–77CrossRefGoogle Scholar
  13. Lakshmanan S, Derin H 1993 Gaussian MRF at multiple resolutions. InMarkov Random Fields — theory and applications (eds) R Chellappa, A Jain (San Diego: Academic Press) ch. 6, pp. 131–158Google Scholar
  14. Li S Z 1995MRF modeling in computer vision (Tokyo: Springer-Verlag)Google Scholar
  15. Mallat S G 1989 Multifrequency channel decompositions of images and wavelet models.IEEE Trans. Acoust. Speech Signal Process. 37: 2091–2110CrossRefGoogle Scholar
  16. Modestino J A, Zhang J 1992 A Markov random field model based approach to image interpretation.IEEE Trans. Pattern Anal. Machine Intell. 14: 606–615CrossRefGoogle Scholar
  17. Nagao M, Matsuyama T 1980A structural analysis of complex aerial photographs (New York: Plenum)Google Scholar
  18. Nanda P K, Sunil Kumar K, Gokhale S, Desai U B 1995 A multiresolution approach to color image restoration and parameter estimation using homotopy continuation method. InProceedings Int. Conf. on Image Processing, Washington, DCGoogle Scholar
  19. Ohta Y 1985Knowledge based interpretation of outdoor natural color scenes (Boston: Pitman)Google Scholar
  20. Peleg S, Federbusch O, Hummel R 1987 Custom made pyramids. InParallel computer vision (ed.) L Uhr (Boston: Academic Press) pp. 125–146Google Scholar
  21. Perez P, Heitz F 1994 Restriction of a MRF on a graph and multiresolution image analysis. Tech. Rep. RR 2170, INRIA, Sophia-Antipolis Cedex, FranceGoogle Scholar
  22. Puliti P, Tascini G 1993 Knowledge-based approach to image interpretation.Image Vision Comput. 11: 122–128CrossRefGoogle Scholar
  23. Rosenfeld A 1984Multiresolution image processing and analysis (Berlin: Springer-Verlag)MATHGoogle Scholar
  24. Schalkoff R J 1989Digital image processing and computer vision (Singapore: John Wiley and Sons)Google Scholar
  25. Sonka M, Tadikonda S K, Collins S M 1993 Genetic algorithms in hypothesize-and-verify image interpretation.Proc. SPIE — Sensor Fusion VI, pp. 236–247Google Scholar
  26. Sunil Kumar K, Desai U B 1994 New algorithms for 3D surface description from binocular stereo using integeration.J. Franklin Inst. B331: 531–554CrossRefGoogle Scholar
  27. Sunil Kumar K, Desai U B 1996 Joint segmentation and image interpretation. Tech. Rep. SPANN.96.2, Indian Institute of Technology — Bombay ( Scholar
  28. Tenenbaum J M, Barrow H G 1977 Experiments in interpretation-guided segmentation.Artif. Intell. 8: 241–274CrossRefGoogle Scholar
  29. Van Essen D C, Anderson C H, Felleman D J 1992 Information processing in the primate visual system: an integrated systems perspective.Science 255: 419–423CrossRefGoogle Scholar

Copyright information

© Indian Academy of Sciences 1998

Authors and Affiliations

  • K Sunil Kumar
    • 1
  • U B Desai
    • 1
  1. 1.Signal Processing and Artificial Neural Networks Laboratory, Department of Electrical EngineeringIndian Institute of Technology — Bombay, PowaiMumbaiIndia

Personalised recommendations